Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 7 additions & 9 deletions connectorx/src/destinations/arrow/arrow_assoc.rs
Original file line number Diff line number Diff line change
Expand Up @@ -288,9 +288,7 @@ fn naive_date_to_arrow(nd: NaiveDate) -> i32 {
}

fn naive_datetime_to_arrow(nd: NaiveDateTime) -> i64 {
nd.and_utc()
.timestamp_nanos_opt()
.unwrap_or_else(|| panic!("out of range DateTime"))
nd.and_utc().timestamp_micros()
}

impl ArrowAssoc for Option<NaiveDate> {
Expand Down Expand Up @@ -328,10 +326,10 @@ impl ArrowAssoc for NaiveDate {
}

impl ArrowAssoc for Option<NaiveDateTime> {
type Builder = TimestampNanosecondBuilder;
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampNanosecondBuilder::with_capacity(nrows)
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: Option<NaiveDateTime>) -> Result<()> {
Expand All @@ -342,17 +340,17 @@ impl ArrowAssoc for Option<NaiveDateTime> {
fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Nanosecond, None),
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
true,
)
}
}

impl ArrowAssoc for NaiveDateTime {
type Builder = TimestampNanosecondBuilder;
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampNanosecondBuilder::with_capacity(nrows)
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: NaiveDateTime) -> Result<()> {
Expand All @@ -363,7 +361,7 @@ impl ArrowAssoc for NaiveDateTime {
fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Nanosecond, None),
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
false,
)
}
Expand Down
152 changes: 143 additions & 9 deletions connectorx/src/destinations/arrowstream/arrow_assoc.rs
Original file line number Diff line number Diff line change
@@ -1,10 +1,11 @@
use super::errors::{ArrowDestinationError, Result};
use super::typesystem::{DateTimeWrapperMicro, NaiveDateTimeWrapperMicro, NaiveTimeWrapperMicro};
use crate::constants::{DEFAULT_ARROW_DECIMAL, DEFAULT_ARROW_DECIMAL_SCALE, SECONDS_IN_DAY};
use crate::utils::decimal_to_i128;
use arrow::array::{
ArrayBuilder, BooleanBuilder, Date32Builder, Date64Builder, Decimal128Builder, Float32Builder,
ArrayBuilder, BooleanBuilder, Date32Builder, Decimal128Builder, Float32Builder,
Float64Builder, Int32Builder, Int64Builder, LargeBinaryBuilder, LargeListBuilder,
StringBuilder, Time64NanosecondBuilder, TimestampNanosecondBuilder, UInt32Builder,
StringBuilder, Time64NanosecondBuilder, Time64MicrosecondBuilder, TimestampNanosecondBuilder, TimestampMicrosecondBuilder, UInt32Builder,
UInt64Builder,
};
use arrow::datatypes::Field;
Expand Down Expand Up @@ -230,6 +231,48 @@ impl ArrowAssoc for Option<DateTime<Utc>> {
}
}

impl ArrowAssoc for DateTimeWrapperMicro {
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampMicrosecondBuilder::with_capacity(nrows).with_timezone("+00:00")
}

#[throws(ArrowDestinationError)]
fn append(builder: &mut Self::Builder, value: DateTimeWrapperMicro) {
builder.append_value(value.0.timestamp_micros());
}

fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, Some("+00:00".into())),
false,
)
}
}

impl ArrowAssoc for Option<DateTimeWrapperMicro> {
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampMicrosecondBuilder::with_capacity(nrows).with_timezone("+00:00")
}

#[throws(ArrowDestinationError)]
fn append(builder: &mut Self::Builder, value: Option<DateTimeWrapperMicro>) {
builder.append_option(value.map(|x| x.0.timestamp_micros()));
}

fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, Some("+00:00".into())),
true,
)
}
}

fn naive_date_to_arrow(nd: NaiveDate) -> i32 {
match nd.and_hms_opt(0, 0, 0) {
Some(dt) => (dt.and_utc().timestamp() / SECONDS_IN_DAY) as i32,
Expand All @@ -238,7 +281,7 @@ fn naive_date_to_arrow(nd: NaiveDate) -> i32 {
}

fn naive_datetime_to_arrow(nd: NaiveDateTime) -> i64 {
nd.and_utc().timestamp_millis()
nd.and_utc().timestamp_micros()
}

impl ArrowAssoc for Option<NaiveDate> {
Expand Down Expand Up @@ -276,10 +319,10 @@ impl ArrowAssoc for NaiveDate {
}

impl ArrowAssoc for Option<NaiveDateTime> {
type Builder = Date64Builder;
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
Date64Builder::with_capacity(nrows)
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: Option<NaiveDateTime>) -> Result<()> {
Expand All @@ -288,15 +331,19 @@ impl ArrowAssoc for Option<NaiveDateTime> {
}

fn field(header: &str) -> Field {
Field::new(header, ArrowDataType::Date64, true)
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
true,
)
}
}

impl ArrowAssoc for NaiveDateTime {
type Builder = Date64Builder;
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
Date64Builder::with_capacity(nrows)
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: NaiveDateTime) -> Result<()> {
Expand All @@ -305,7 +352,56 @@ impl ArrowAssoc for NaiveDateTime {
}

fn field(header: &str) -> Field {
Field::new(header, ArrowDataType::Date64, false)
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
false,
)
}
}

impl ArrowAssoc for Option<NaiveDateTimeWrapperMicro> {
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: Option<NaiveDateTimeWrapperMicro>) -> Result<()> {
builder.append_option(match value {
Some(v) => Some(v.0.and_utc().timestamp_micros()),
None => None,
});
Ok(())
}

fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
true,
)
}
}

impl ArrowAssoc for NaiveDateTimeWrapperMicro {
type Builder = TimestampMicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
TimestampMicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: NaiveDateTimeWrapperMicro) -> Result<()> {
builder.append_value(value.0.and_utc().timestamp_micros());
Ok(())
}

fn field(header: &str) -> Field {
Field::new(
header,
ArrowDataType::Timestamp(TimeUnit::Microsecond, None),
false,
)
}
}

Expand Down Expand Up @@ -349,6 +445,44 @@ impl ArrowAssoc for NaiveTime {
}
}

impl ArrowAssoc for Option<NaiveTimeWrapperMicro> {
type Builder = Time64MicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
Time64MicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: Option<NaiveTimeWrapperMicro>) -> Result<()> {
builder.append_option(value.map(|t| {
t.0.num_seconds_from_midnight() as i64 * 1_000_000 + (t.0.nanosecond() as i64) / 1000
}));
Ok(())
}

fn field(header: &str) -> Field {
Field::new(header, ArrowDataType::Time64(TimeUnit::Microsecond), true)
}
}

impl ArrowAssoc for NaiveTimeWrapperMicro {
type Builder = Time64MicrosecondBuilder;

fn builder(nrows: usize) -> Self::Builder {
Time64MicrosecondBuilder::with_capacity(nrows)
}

fn append(builder: &mut Self::Builder, value: NaiveTimeWrapperMicro) -> Result<()> {
builder.append_value(
value.0.num_seconds_from_midnight() as i64 * 1_000_000 + (value.0.nanosecond() as i64) / 1000,
);
Ok(())
}

fn field(header: &str) -> Field {
Field::new(header, ArrowDataType::Time64(TimeUnit::Microsecond), false)
}
}

impl ArrowAssoc for Option<Vec<u8>> {
type Builder = LargeBinaryBuilder;

Expand Down
15 changes: 15 additions & 0 deletions connectorx/src/destinations/arrowstream/typesystem.rs
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,15 @@ use crate::impl_typesystem;
use chrono::{DateTime, NaiveDate, NaiveDateTime, NaiveTime, Utc};
use rust_decimal::Decimal;

#[derive(Debug, Clone, Copy)]
pub struct DateTimeWrapperMicro(pub DateTime<Utc>);

#[derive(Debug, Clone, Copy)]
pub struct NaiveTimeWrapperMicro(pub NaiveTime);

#[derive(Debug, Clone, Copy)]
pub struct NaiveDateTimeWrapperMicro(pub NaiveDateTime);

#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, PartialOrd, Ord)]
pub enum ArrowTypeSystem {
Int32(bool),
Expand All @@ -16,8 +25,11 @@ pub enum ArrowTypeSystem {
LargeBinary(bool),
Date32(bool),
Date64(bool),
Date64Micro(bool),
Time64(bool),
Time64Micro(bool),
DateTimeTz(bool),
DateTimeTzMicro(bool),
Float32Array(bool),
}

Expand All @@ -36,8 +48,11 @@ impl_typesystem! {
{ LargeBinary => Vec<u8> }
{ Date32 => NaiveDate }
{ Date64 => NaiveDateTime }
{ Date64Micro => NaiveDateTimeWrapperMicro }
{ Time64 => NaiveTime }
{ Time64Micro => NaiveTimeWrapperMicro }
{ DateTimeTz => DateTime<Utc> }
{ DateTimeTzMicro => DateTimeWrapperMicro }
{ Float32Array => Vec<Option<f32>> }
}
}
43 changes: 28 additions & 15 deletions connectorx/src/transports/oracle_arrowstream.rs
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
use crate::{
destinations::arrowstream::{
typesystem::ArrowTypeSystem, ArrowDestination, ArrowDestinationError,
typesystem::{ArrowTypeSystem, DateTimeWrapperMicro, NaiveDateTimeWrapperMicro}, ArrowDestination, ArrowDestinationError,
},
impl_transport,
sources::oracle::{OracleSource, OracleSourceError, OracleTypeSystem},
Expand Down Expand Up @@ -29,19 +29,32 @@ impl_transport!(
systems = OracleTypeSystem => ArrowTypeSystem,
route = OracleSource => ArrowDestination,
mappings = {
{ NumFloat[f64] => Float64[f64] | conversion auto }
{ Float[f64] => Float64[f64] | conversion none }
{ BinaryFloat[f64] => Float64[f64] | conversion none }
{ BinaryDouble[f64] => Float64[f64] | conversion none }
{ NumInt[i64] => Int64[i64] | conversion auto }
{ Blob[Vec<u8>] => LargeBinary[Vec<u8>] | conversion auto }
{ Clob[String] => LargeUtf8[String] | conversion none }
{ VarChar[String] => LargeUtf8[String] | conversion auto }
{ Char[String] => LargeUtf8[String] | conversion none }
{ NVarChar[String] => LargeUtf8[String] | conversion none }
{ NChar[String] => LargeUtf8[String] | conversion none }
{ Date[NaiveDateTime] => Date64[NaiveDateTime] | conversion auto }
{ Timestamp[NaiveDateTime] => Date64[NaiveDateTime] | conversion none }
{ TimestampTz[DateTime<Utc>] => DateTimeTz[DateTime<Utc>] | conversion auto }
{ NumFloat[f64] => Float64[f64] | conversion auto }
{ Float[f64] => Float64[f64] | conversion none }
{ BinaryFloat[f64] => Float64[f64] | conversion none }
{ BinaryDouble[f64] => Float64[f64] | conversion none }
{ NumInt[i64] => Int64[i64] | conversion auto }
{ Blob[Vec<u8>] => LargeBinary[Vec<u8>] | conversion auto }
{ Clob[String] => LargeUtf8[String] | conversion none }
{ VarChar[String] => LargeUtf8[String] | conversion auto }
{ Char[String] => LargeUtf8[String] | conversion none }
{ NVarChar[String] => LargeUtf8[String] | conversion none }
{ NChar[String] => LargeUtf8[String] | conversion none }
{ Date[NaiveDateTime] => Date64Micro[NaiveDateTimeWrapperMicro] | conversion option }
{ Timestamp[NaiveDateTime] => Date64Micro[NaiveDateTimeWrapperMicro] | conversion none }
{ TimestampNano[NaiveDateTime] => Date64[NaiveDateTime] | conversion auto }
{ TimestampTz[DateTime<Utc>] => DateTimeTzMicro[DateTimeWrapperMicro] | conversion option }
{ TimestampTzNano[DateTime<Utc>] => DateTimeTz[DateTime<Utc>] | conversion auto }
}
);
impl TypeConversion<NaiveDateTime, NaiveDateTimeWrapperMicro> for OracleArrowTransport {
fn convert(val: NaiveDateTime) -> NaiveDateTimeWrapperMicro {
NaiveDateTimeWrapperMicro(val)
}
}

impl TypeConversion<DateTime<Utc>, DateTimeWrapperMicro> for OracleArrowTransport {
fn convert(val: DateTime<Utc>) -> DateTimeWrapperMicro {
DateTimeWrapperMicro(val)
}
}
Loading